GitHub - ComfyNodePRs/PR-ComfyUI-HunyuanImageLatentToVideoLatent-abd992c9: A ComfyUI node which copies a given latent's samples tensor along the time axis ((length - 1) // 4) + 1 times to form a longer latent and then prepares a noise_mask tensor of the same shape such that the value of the mask for a given time step is given by the function at https://www.desmos.com/calculator/vhw74mr1vh.

ComfyNodePRs / PR-ComfyUI-HunyuanImageLatentToVideoLatent-abd992c9 Public

forked from philiprodriguez/ComfyUI-HunyuanImageLatentToVideoLatent

Notifications You must be signed in to change notification settings
Fork 0
Star 0

A ComfyUI node which copies a given latent's samples tensor along the time axis ((length - 1) // 4) + 1 times to form a longer latent and then prepares a noise_mask tensor of the same shape such that the value of the mask for a given time step is given by the function at https://www.desmos.com/calculator/vhw74mr1vh.

MIT license